Recent clinical data demonstrates a clear dose-tumor response relationship for both conventional fractionation and hypofractionated nonsmall-cell lung cancer radiotherapy; several large studies confirm the dose-toxicity relationship for lung, esophagus, heart and bronchus. However, lung tumors move with respiration, change position with respect to skeletal anatomy from day to day, and generally decrease in volume during a course of radiotherapy. These anatomic changes hinder the accurate imaging, planning and delivery of thoracic radiotherapy and, hence, impede our ability to maximize the therapeutic window between local control/survival and treatment-induced complications. Developments in respiratory gating, breath-hold radiotherapy and four-dimensional (4D) radiotherapy account for some of the respiratory motion issues, however, these techniques have introduced additional problems that have yet to be addressed. For example, respiratory irregularities in the 4D CT acquisition process manifest themselves as image artifacts and subsequently as systematic errors throughout the entire treatment course. Temporal variations between anatomy motion and the respiratory signal used as a surrogate for this motion introduce significant errors, since the correlation between the surrogate and anatomy motion is location-dependent, and anatomy and surrogate motions vary from cycle to cycle and day today. The overall goal of the project is to systematically quantify anatomic variations throughout a course of radiotherapy for a representative patient cohort, develop strategies to mitigate the effect of the variations, and, thereby, minimize their clinical impact whilst rigorously accounting for the residual uncertainties. To achieve the goal the specific aimsare: (1) To conduct a clinical imaging study to quantify the magnitude and distribution of inter- and intrafraction anatomic variations, including the temporal stability of the tumor/respiration signal correlation. (2) To improve the acquisition and reconstruction of 4D CT images by (a) advancing the 4D CT data collection process and (b) evaluating 4D CT image reconstruction using different respiratory inputs, in addition to quantifying the uncertainty of deformable image-registration algorithms, a core tool for 4D IGART. (3) To develop and investigate the efficacy of inter- and intrafraction probabilistic-based 4D IGART strategies for clinical application and to perform simulation studies to determine the dosimetric and biological improvement gains from the 4D IGART system. The culmination of the project is the clinical implementation of 4D IGART safety and efficacy studies.